Download - A Vegetation Index of Biotic
-
7/28/2019 A Vegetation Index of Biotic
1/69
A vegetation index of bioticintegrity for small-order streams in
southwest Montana and a floristic
quality assessment for westernMontana wetlands.
Prepared for:
Montana Department of Environmental Quality
and
U.S. Environmental Protection Agency
By:
W. M. Jones
Montana Natural Heritage Program
Natural Resource Information System
Montana State Library
August 2005
-
7/28/2019 A Vegetation Index of Biotic
2/69
-
7/28/2019 A Vegetation Index of Biotic
3/69
A vegetation index of bioticintegrity for small-order streams in
southwest Montana and a floristic
quality assessment for westernMontana wetlands.
Prepared for:
Montana Department of Environmental Quality
and
U.S. Environmental Protection Agency
Agreement Number:
203097
By:
W. M. Jones
-
7/28/2019 A Vegetation Index of Biotic
4/69
-
7/28/2019 A Vegetation Index of Biotic
5/69
ABSTRACT
This study evaluated the relationship betweengrazing-related disturbances and vegetation in first-
through third-order montane streams in
southwestern Montana. Eight vegetation metrics
(relative cover of native graminoids, relative cover
of exotic species, relative cover of hydrophytes,
cover-weighted floristic quality index, cover-
weighted mean bank stability rating, absolute
combined cover of seedling and young willows, and
willow seedling density) were found to respond to
grazing-related disturbances. These metrics were
combined into a multimetric index, the vegetation
index of biotic integrity (VIBI), which respondedstrongly to a grazing-associated disturbance
gradient. VIBI scoring thresholds were established
that differentiated among three condition classes:
reference condition, moderately impaired, and
severely impaired. The VIBI can be used as an
evaluation tool to assess riparian area condition.
Coefficients of conservatism, which form the basis
for floristic quality assessments, were assigned by
an expert panel for plant species likely to occur in
western Montana wetlands.
-
7/28/2019 A Vegetation Index of Biotic
6/69
ACKNOWLEDGMENTS
This study was funded through a U.S.Environmental Protection Agency wetland
protection grant administered by the Montana
Department of Environmental Quality. My sincere
thanks to all who assisted with this project: Lynda
Saul, for her tireless efforts as the linchpin for
wetland conservation at Montana DEQ; Randy
Apfelbeck, Anna Noson, and Bryce Maxell, for
their fruitful involvement with the wetland
monitoring and assessment work group; SteveCooper, for his review of riparian assessment
methods; and Greg Kudray and Coburn Currier for
editing and formatting the final version of this
report. My highest thanks also to Peter Lesica,
John Pierce, Steve Shelly, Mary Manning, Steve
Cooper, and Scott Mincemoyer for their invaluable
help in assigning coefficients of conservation to
western Montana wetland plants.
-
7/28/2019 A Vegetation Index of Biotic
7/69
TABLEOFCONTENTS
Introduction ................................................................................................................................................... 1Methods ........................................................................................................................................................ 3
Study Area .............................................................................................................................................. 3
Site Selection .......................................................................................................................................... 3
Data Collection ...................................................................................................................................... 3
Human Disturbance Gradient ................................................................................................................ 6
Multimetric Analysis ............................................................................................................................... 7
Whole Community Analysis ................................................................................................................. 11
Spatial Autocorrelation Analysis........................................................................................................... 12
Results ........................................................................................................................................................ 13
Human Disturbance Gradient .............................................................................................................. 13
Metrics ................................................................................................................................................. 14
VIBI ..................................................................................................................................................... 15
Whole Community................................................................................................................................ 20
Spatial Autocorrelation ......................................................................................................................... 20
Discussion ................................................................................................................................................... 21
Recommendations for Future Improvements ...................................................................................... 22
Literature Cited .......................................................................................................................................... 24
Appendix A. Coefficients of conservation for selected wetland plants that occur in western Montana.
Appendix B. List and attributes of sampled plant species.
Appendix C. Location and condition rating of sample reaches.
LISTOFFIGURES
Figure 1. Location and functional condition classes of sample reaches ...................................................... 4
Figure 2. Schematic showing placement of and data collected for subsamples within sample reaches ..... 5
Figure 3. Graphical representation of the relationship between PFC categories and the
composite disturbance gradient .................................................................................................. 13
Figure 4. Discriminatory power of selected metrics and their relationship with a composite
human disturbance gradient ................................................................................................... 16-17
Figure 5. Scatterplot showing the relationship between the vegetation index of biotic integrity
(VIBI) and a composite human disturbance gradient................................................................. 15Figure 6. Reference condition site ............................................................................................................. 18
Figure 7. Moderately impaired site ............................................................................................................ 18
Figure 8. Severely impaired site ................................................................................................................. 18
Figure 9. Tree diagram showing VIBI scoring thresholds associated with disturbance categories
and scatterplot of the composite disturbance gradient and VIBI ............................................... 19
-
7/28/2019 A Vegetation Index of Biotic
8/69
LISTOFTABLES (CONTINUED)
Table 4. Candidate metrics considered for inclusion in the VIBI .............................................................. 14Table 5. Formulas used to score metrics ................................................................................................... 15
Table 6. Accuracy assessment of VIBI scoring thresholds with regard to disturbance classes ............... 18
Table 7. Species indicative of reference, moderately disturbed, and severely disturbed sites................... 19
-
7/28/2019 A Vegetation Index of Biotic
9/69
INTRODUCTIONThe list of economic and environmental benefits
provided by wetlands and riparian areas is long.These benefits include groundwater recharge,filtration and storage of sediments, nutrients, andpollutants, floodwater storage and attenuation, andunique habitat values (Brinson et al. 1981, Keddy2000). Consequently, the importance of wetlandsand riparian areas is disproportionate to theirphysical extent on the landscape, especially in
semiarid regions such as Montana (Finch andRuggiero 1993, Patten 1998). Despite theirimportance to both humans and wildlife, anestimated 25% of Montanas wetlands have beenlost in the past 200 years (Dahl 1990). To improvewetland conservation in Montana, the MontanaDepartment of Environmental Quality (DEQ) isdeveloping a comprehensive statewide wetland
monitoring and assessment program, of which thispresent study is a part.
This program will use a three-tiered approach tocharacterize the condition and extent of wetlands inMontana. DEQ will combine landscape-levelremotely sensed data, rapid site-level assessments,and detailed site-level evaluations of biota toevaluate wetland condition and to identify
anthropogenic stressors that limit that condition.The purpose of the present study was to identifyattributes of the riparian vegetation community ofsmall-order streams that responded predictably tohuman disturbance. Such attributes could then beused as indicators of wetland condition for detailedsite assessments as well as for validating andcalibrating rapid assessment methods.
I used a multimetric approach to identifyvegetation indicators. Multimetric analysisattempts to determine the status of a wetland orstream reach by directly measuring the condition ofone or more of its biotic components (Danielson2002). This method is based on defining a
l ti l h t d i t d
easy to measure and interpret (Karr and Chu
1999). Successful metrics can then be combinedinto a multimetric index that reflects a diverse bioticresponse to human-related stressors and is anintegrative measure of the sites biological condition(Karr and Chu 1999, Teels and Adamus 2002).
Biological assessments can be accurate andcost-effective tools to assess wetland and streamcondition and to measure impairment (Karr and
Chu 1999). Since biota integrate multiple physicaland chemical parameters, directly measuring abiotic communitys response to anthropogenicstressors can be the most effective means toevaluate the effect of those stressors on wetlandcondition and function (Danielson 2002). The utilityof using biota to assess wetlands has beendemonstrated for numerous taxa, including fish
(Karr 1981, Hughes et al. 1998, Mebane et al.2003), diatoms (Fore and Grafe 2002), benthic andterrestrial macroinvertebrates (Kimberling et al.2001, Blocksom et al. 2002, Klemm et al. 2003),birds (Bryce et al. 2002), and vegetation(DeKeyser et al. 2003, Mack 2004, Ferreira et al.2005). This approach has been shown to beeffective for perennial and seasonal depressional
wetlands and ephemeral and intermittent streams inMontana (Apfelbeck 2001, Jones 2004).This study was conducted in southwestern
Montana where the primary human-relatedstressors are livestock grazing and agriculture.Livestock grazing can influence numerous physicalparameters in riparian systems, including streamchannel and bank geomorphology and stability
(Kauffman et al. 1983b, Clary 1999, Clary andKinney 2002), floodplain microchannel sinuosityand drainage density (Flenniken et al. 2001), andsoil bulk density, pore space, infiltration, andpotential nitrification and mineralization rates(Kauffman and Krueger 1984, Wheeler et al. 2002,K ff t l 2004) B lt i th h i l
-
7/28/2019 A Vegetation Index of Biotic
10/69
2002, Thorne et al. 2005), and increase theabundance of weedy species, such as Kentuckybluegrass (Poapratensis L.) (Schulz and Leininger1990, Green and Kauffman 1995), possibly byaltering competitive interactions with nativegraminoids (Martin and Chambers 2001). At lower
elevations, agricultural land uses and theirassociated hydrologic modifications becomeimportant stressors on riparian systems; however,this study was conducted on smaller order streamsthat were largely unaffected by agriculturaldisturbances.
-
7/28/2019 A Vegetation Index of Biotic
11/69
METHODS
Study AreaThe study area encompassed portions of
Beaverhead and Madison Counties in southwestMontana (Figure 1). This area lies within theNorthern Rocky Mountain and Montana Valley andFoothill Prairies Ecoregions (Woods et al. 1999)and is characterized by broad intermontane valleysinterspersed with isolated mountain ranges. Thegeology is a complex mixture of predominatelyTertiary and Cretaceous sedimentary rocks withlocalized intrusions of Tertiary volcanics,Mississippian limestone, Proterozoic quartzite, andArchaean gneiss and schist; Pleistocene glacialdeposits are locally abundant at higher elevations(Ruppel et al. 1993, Ruppel 1999, Lonn et al. 2000,Skipp and Janecke 2004). The climate is semiaridand continental. The weather station at Lima,Montana, which is representative of lowerelevation sample locations, has recorded meantemperatures ranging from 16.8F in January to61.3F in July and mean precipitation of 12.43inches annually (Western Regional Climate Center2005).
Site Selection
Potential sample locations were limited to small-order streams that had been previously evaluatedfor functional status by the Bureau of LandManagement (BLM) and U.S. Forest Service(USFS) using standardized riparian assessments.
BLM assessments used the proper functioningcondition (PFC) methodology, which combinesqualitative evaluations of hydrology, vegetation, anderosion/deposition to evaluate a stream reach(Prichard et al. 1998). USFS assessmentsevaluated a stream reachs degree of departurefrom reference condition using quantitative hydro
stratified by condition class. Rated reaches were
displayed in a geographic information system(ArcGIS 8.3, ESRI, Redlands, California), and 11functioning, 9 functioning at risk, and 10nonfunctioning reaches were selected. All 30stream reaches were sampled from June to August2004.
Sample reaches were first- through third-order,low gradient streams ranging in elevation from
6,000 to 7,900 feet above sea level; most reacheswould be categorized as E type streams underRosgens (1996) classification system. All samplereaches were on tributaries to the Beaverhead andRed Rock Rivers on lands managed by the BLM orUSFS and supported varying levels of willow cover,predominately Geyers willow (Salix geyerianaAnderss.), Booths willow (S. boothii Dorn), and
Drummonds willow (S. drummondiana Barratt exHook.). Dominant herbaceous species includedbeaked sedge (Carex utriculata Boott), watersedge (C.aquatilis Wahlenb.), Baltic rush (Juncusbalticus Willd.), bluejoint reedgrass(Calamagrostis canadensis (Michx.) Beauv.), andKentucky bluegrass.
Data Collection
The sampling method used to collect speciesabundance and environmental data was modifiedfrom the techniques outlined in Winward (2000)and Coles-Ritchie et al. (2004) and was selectedbased in part on a review by Cooper (2004). Thesample unit was a 100-m stream reach that was
subsampled using two types of systematicallyplaced sample frames: 0.1-m2 (0.2-m 0.5-m)quadrats and 4-m2 (1.13-m radius) plots. Sampleframes were placed along transects runningperpendicular and parallel to the stream channel,such that an area of 100-m 8-m was sampled
-
7/28/2019 A Vegetation Index of Biotic
12/69
4
Figure 1. Locations and functional condition classes of sample reaches. Condition classes are PFC (proper functioning condition), FAR (functioning at risk),
and NF (nonfunctioning)
-
7/28/2019 A Vegetation Index of Biotic
13/69
and 10 plots placed at 10-m intervals, with groupsof four quadrats and one plot being placed onalternating sides of the channel, with the long endsof quadrats placed parallel to the channel. Fivetransects were also placed perpendicular to thevalley slope at an interval of 20 m on alternatingsides of the channel. Three quadrats (located 2.5,
5.0, and 7.5 m from the greenline with long endsparallel to the transect) and one plot (located 5.0 mfrom the greenline) were sampled at each transect.
Species abundances were recorded using thecover estimation method described by Daubenmire(1959). Six cover classes were used to recordspecies abundances: 1 (95% cover). Herbaceousvegetation was sampled in quadrats and woodyvegetation was sampled in plots. For woodyspecies, both total cover and cover by age class(Table 1) were estimated. Mean height for eachage class was estimated to the nearest 0.1 m. The
number of woody seedlings present in each plotwas also recorded. Species nomenclature followsthe PLANTS database (version 3.5), which is thenational naming standard used by the federalgovernment (Natural Resources ConservationService 2004).
Five potential indicators of grazing-related
stressors were measured: (1) amount of bareground, (2) number of hoof shears (pugs) present ineach plot, (3) number and mean depth ofhummocks present in each plot, (4) bank stability atgreenline plots, and (5) browse intensity. Bareground was measured as the number of quadratcorners that intersected bare mineral soil. Bankstability was evaluated with a 0.15-m wide plot
running from the scour line to either twice bankfullheight or a flat depositional surface, whichever waslower. A bank was considered unstable if less than50% of the plot was covered by perennialvegetation ground cover or roots, rocks greaterthan 0.15-m diameter, or logs greater than 0.1-m
Figure 2. Schematic showing placement of and data collected for subsamples within sample reaches.
0.2-m 0.5-m quadrat (abundance of herbaceous vegetation, bare ground, height above bankfulldischarge)
4-m circular plot (abundance of woody vegetation, pugging/hummocking density, bankstability, browse intensity)
8 m
8 m100-m Sample Reach
-
7/28/2019 A Vegetation Index of Biotic
14/69
diameter, either singly or in combination. Browseintensity was evaluated following the method ofKeigley and Frisina (1998). The plant nearest theplot center with a primary stem between 0.5 and1.5 m high was selected for evaluation; if only tallerplants were present, the plant nearest the plotcenter that had a primary stem with a terminalleader within the browse zone (0.5 to 1.5 m high)that was not mechanically protected was selected.If browsing had killed an entire annual segment onthe stem selected, browse intensity was considered
heavy; if not, browse intensity was considered lightto moderate. This evaluation was performed onlyif a palatable species with a terminal leader withinthe browse zone was present (e.g., Salix spp.,Cornus sericea L., Populus tremuloides Michx.).Finally, the elevation of greenline quadrats inrelation to bankfull discharge was measured to thenearest 0.01 m.
To determine whether the number of subsampleswas adequately characterizing vegetation, Iexamined species area curves for the initial fivesites surveyed to see if they met the criterion ofless than a 5% increase in the number of speciessampled for a 10% increase in sample area(Mueller-Dombois and Ellenberg 1974). All fivesites met this threshold by the time 80% of the area
had been surveyed; therefore, the samplingintensity was considered adequate.
All data were aggregated to the level of samplereach. Vegetation abundance was calculated byaveraging cover class midpoints. Values for allother variables were averaged except for browseintensity and bank stability, which were calculatedas frequencies.
Human Disturbance Gradient
Disturbance parameters measured on-site werechosen to be responsive to grazing-related stresses.Two other factors were calculated: allotment
sink-filled digital elevation model (30-m rasterNational Elevation Dataset, US Geological Survey2002). Another commonly used disturbanceindicator, percent of catchment in agricultural orother human-modified land cover, was notconsidered because catchment land cover wascomprised almost entirely of native vegetation forall sites. The interpretation of this measurementhas also been shown to be problematic because ofspatial autocorrelation among land cover classes(King et al. 2005).
To rank sites by their overall disturbance, Icalculated a composite disturbance measure usingprincipal components analysis (PCA). PCAidentifies linear combinations of the variables thatexplain the greatest variation in the data. Theoriginal dataset can thereby be represented infewer dimensions with composite variables, termedprincipal components, than the number of original
descriptors. Although PCA was developed for datawith multivariate normal distributions, it is robust todepartures from normality as long as factors arerelatively unskewed (Legendre and Legendre1998). Disturbance factors were transformed tomeet the threshold recommended by McCune andGrace (2002) of |skewness|
-
7/28/2019 A Vegetation Index of Biotic
15/69
disturbance scores were different among PFC
categories. To compensate for multiple testing,
significance values were modified with a
Bonferroni correction (Sokal and Rohlf 1995).
Assumptions of analysis of variance (normal
distribution of residuals and homogenous error
variances) were examined graphically and with
Levenes test (Levene 1960). Because BLM and
USFS methodologies differed somewhat, I tested
all sites (n = 30) and sites on BLM land (n = 23).
Sites on USFS land were not tested independently
due to small sample size (n = 7).
Multimetric Analysis
Vegetation response to the composite
disturbance gradient was quantified by developing a
multimetric index, termed the vegetation index of
biotic integrity (VIBI). VIBI development includedseveral steps. Candidate metrics were screened
for their ability to discriminate between least and
most disturbed sites and for their overall response
to the disturbance gradient. Metrics found to be
responsive to human disturbance were tested for
redundancy and the most responsive, non-
redundant metrics were combined into the VIBI.
Metric and VIBI analyses were conducted using R
statistical software, except where otherwise noted.
Candidate Metrics
Vegetation attributes that have been considered
in other studies fall into several categories:
community-based metrics (e.g., species richness
and dominance), metrics based on plant functional
groups (e.g., annuals, perennials, disturbance-tolerant species), and species-specific metrics
(Fennessy et al. 2002). Potential metrics
considered in this study were largely derived from
the second category. Among those considered
were metrics that had been proven to be effective
i i t di (B th 1998 H l d G
to decrease with increasing disturbance were the
relative cover of native perennials, native
graminoids, and Carices. Those expected to
increase with disturbance included the relative
cover of exotic species, exotic grasses, and
annuals/biennials. Several metrics related to the
woody vegetation component were also considered,
including density of willow seedlings, cover of
willow seedlings, cover of young willows, combined
cover of willow seedlings and young willows, total
willow cover, and willow age distribution, which
was calculated as the combined cover of willow
seedlings and young willows divided by total willow
cover. Cover values for willow-related metrics
were calculated with absolute cover values to
emphasize structural differences among sites. All
were expected to decrease with human
disturbance.
Metrics based on diversity measures Twodiversity measures, the Shannon index and the
reciprocal of the Simpson dominance index, were
calculated. Both indices are related to and based
partly on species richness; however, they also
incorporate the equitability of species abundances
as well. For example, both indices would rate a
plot with one dominant and two incidental species
as less diverse than a plot with three equallyabundant species. Shannon diversity is calculated
as
H = - pilogp
i
where H is the Shannon diversity index andpiis
the relative cover of species i within a sample unit.
Simpson diversity is calculated as
D = 1 / pi
2
where D is the Simpson diversity index. These two
measures are similar but vary in their sensitivity to
rare species, with Simpson diversity being
-
7/28/2019 A Vegetation Index of Biotic
16/69
Swink and Wilhelm (1979). Their floristic quality
assessment index (FQAI) is based on the
perceived affinity of native plant species to
particular habitats and their tolerance to
disturbance. Within a regional flora, each species
affinity/tolerance is subjectively quantified using an
11-point ordinal standard termed the coefficient of
conservatism (C) (Table 2). C-values are assigned
by an expert panel of botanists familiar with the
flora in question. Using C-values, the floristic
quality, and by extension, condition, of different
sites can be compared.
The FQAI and derived measures have proven to
be highly sensitive indicators of disturbance. The
usefulness of the FQAI as a vegetation metric has
been demonstrated for prairie potholes and
ephemeral streams in the Great Plains (Mushet et
al. 2002, DeKeyser et al. 2003, Jones 2004),
depressional wetlands in Florida (Cohen et al.
2004), numerous wetland types in Ohio (Lopez andFennessy 2002, Andreas et al. 2004), and
woodlands in southern Ontario (Francis et al.
2000).
Computationally, the FQAI is based on the mean
C-value of a sites vegetation, which is calculated
as:
mean Cj
= Cij/ n
j
where Cij
is the coefficient of conservatism of
native species i at sitej and nj
is the number of
native species at sitej. The FQAI score for sitej
is calculated as:
FQAIj
= Cij/ n
j= (mean C
j) n
j
The square root modifier was proposed by Wilhelm
and Ladd (1988) to dampen the effects of species
richness on the index. This diminishes disparities
between high quality species-poor sites and lower
quality species-rich sites. C-values used in this
study were determined by a panel of expert
botanists and are listed in Appendix A.
I applied two modifications to the standard FQAI
method similar to those proposed by Cohen et al.
(2004). First, I included exotic species in the
calculation of the FQAI, which are typically not
considered. However, exotic species are an
important indicator of site quality and their inclusion
in the index seems warranted. The other
modification was to weight each species C-value
by its relative abundance. Abundance is a more
sensitive measure of species response than
presence-absence (Rahel 1990), so it is possible
that a cover-weighted FQAI may be a better
indicator of site condition than the standard
formulation. The cover-weighted FQAI (cFQAI)
for sitej was calculated as:
cFQAIj
= [ (Cijaij) / aij] nj
where aij
is the abundance (measured as cover) of
species i at sitej and Cij
is the coefficient of
conservatism of species i at sitej. Exotic species
were included in the calculation of the cover-
weighted FQAI. All calculations of mean C-value
and FQAI were expected to decrease withdisturbance.
In addition to the FQAI itself, there are several
other potential metrics that can be derived from C-
values. These include the relative cover of
disturbance-tolerant species (species with C-values
< 2) and disturbance-intolerant species (species
with C-values > 6).
Metrics based on wetland indicator status
Wetland indicator status is a reflection of a species
affinity for wetland habitats. Species are placed
into one of five ordinal categories that represent the
Table 2. Coefficient of conservatism scoring criteria (afterAndreas et al. 2004).
/
/
-
7/28/2019 A Vegetation Index of Biotic
17/69
likelihood of its occurring in wetlands versus non-
wetlands. These categories, scored one through
five, are: 1 = obligate upland (species occur almost
exclusively in uplands), 2 = facultative upland
(species usually occur in non-wetlands), 3 =
facultative (species equally likely to occur in
wetlands or non-wetlands), 4 = facultative wetland
(species usually occur in wetlands), and 5 =
obligate wetland (species occur almost exclusively
in wetlands). Indicator status values were obtained
from the 1988 national list and 1993 Pacific
Northwest supplement published by the U.S. Fishand Wildlife Service (Reed 1993). Indicator values
for the Pacific Northwest (Region 9) were used.
These lists only identified obligate upland species if
they occurred in wetlands in another region.
Species sampled in this study that did not occur on
the lists were coded as obligate upland species.
Three potential metrics were calculated from
wetland indicator values: relative cover ofhydrophytes (species with an indicator value of
obligate or facultative wetland), relative cover of
upland species (species with an indicator value of
obligate or facultative upland), and the cover-
weighted mean wetland indicator value, which is
calculated as:
cWIj = (WIijaij) / aij
where cWIj
is the cover-weighted mean wetland
indicator value for sitej, WIij
is the wetland
indicator value of species i at sitej, and aij
is the
abundance of species i at sitej. Relative cover of
hydrophytes and cWI were expected to decline
with increasing disturbance, while the relative
cover of upland-associated species was expectedto increase.
Metrics based on bank stability rating The
last category of metrics were derived from the
ability of species to stabilize streambanks either
potential metrics were calculated: relative cover of
stabilizing species (species with stability ratings of
good or excellent) and the cover-weighted mean
bank stability rating, which was calculated as:
cSRj
= (SRija
ij) / a
ij
where cSRj
is the cover-weighted mean vegetation
stability rating for sitej, SRij
is the stability rating of
species i at sitej, and aij
is the abundance of
species i at sitej. Only data from greenline
transects were used to calculate bank stabilitymetrics. Both metrics were expected to decrease
with disturbance. Stability ratings for species are
listed in Appendix B.
Metric Evaluation and Selection
A three-step selection process was used to
evaluate candidate metrics for inclusion in the
VIBI, similar to Blocksom et al. (2002). The three
criteria were the ability of metrics to discriminate
between least and most disturbed sites, the overall
relationship between metrics and the composite
disturbance gradient, and redundancy among
metrics. To test discriminatory power, I identified
least disturbed sites (disturbance score 75th percentile ofdisturbance index). Percentiles were calculated in
the R statistical package using the method
recommended by Hyndman and Fan (1996). Box
plots were used to examine metric distributions.
Metrics were scored based on their ability to
differentiate between the two disturbance
categories using the methodology described by
Barbour et al. (1996). Metrics that had no overlapof interquartile range (middle 50% of observations)
were scored 3, those that had no overlap of median
and interquartile range were scored 2, those that
had an overlap of one median and interquartile
range were scored 1, and those where both
-
7/28/2019 A Vegetation Index of Biotic
18/69
Finally, to ensure that metrics would not beproviding redundant information to the VIBI, Iexamined correlations among the remainingcandidate metrics. I used the high thresholdrecommended by the U.S. EnvironmentalProtection Agency (|r
s| >0.9) to determine
redundancy (USEPA 1998). Where two or moremetrics were found to be redundant, the one withthe greatest discriminatory power and greatestresponse to disturbance was retained.
Metric ScoringMetrics are usually scored by assigning value
ranges to discrete categories depending on theirdeviation from an expected reference condition(Karr 1981, Wilcox et al. 2002, DeKeyser et al.2003, Mack 2004). A commonly used scheme is toassign reference condition sites a score of 5, sitesthat deviate somewhat from reference condition ascore of 3, and sites that strongly deviate fromreference condition a score of 1 (Karr and Chu1999). However, others have suggested thatscoring metrics along a continuous scale would bemore accurate, less variable, and easier to interpret(Minns et al. 1994, Hughes et al. 1998, McCormicket al. 2001, Mebane et al. 2003). Blocksom (2003)found that continuous scoring improved the overall
performance of the multimetric index whencompared to discrete scoring methods.
Before scoring metrics I first identified the 95 th
percentile value of each metric (5 th percentile valueof metrics that increased in response todisturbance), which I used as the best expectedvalue to reduce the effect of outliers (Barbour etal. 1999). Metrics were scored by linear
interpolation. Scores of metrics that decreased inresponse to disturbance were calculated by dividingthe observed value by the 95th percentile value;scores of metrics that increased in response todisturbance were calculated by dividing thedifference between the maximum and observed
disturbance gradient were log-transformed prior toscoring to improve linearity in their response to thecomposite disturbance gradient. Logtransformations were chosen based on the Box-Cox power transformation constrained by thedisturbance gradient. The Box-Cox parameter wasestimated using the MASS package (Venables andRipley 2002) for R software.
VIBI Scoring and EvaluationVIBI scores were calculated by averaging
scores of selected metrics and multiplying by 100.The VIBI therefore ranged from 0 to 100regardless of the number of metrics found to beinterpretable. The strength of the relationshipbetween the VIBI and the composite disturbancegradient was evaluated using ordinary least squaresregression. Assumptions of linear regression(normal distribution, constant variance, andindependence of errors) were examinedgraphically.
One application of the VIBI is to use it as avalidation tool to assess the accuracy of rapidassessments. The output of the rapid assessmentis an ordinal rating of wetland condition. Toprovide a congruent VIBI scoring system, I wantedto determine how many condition classes the VIBI
could accurately distinguish and to identify scoringthresholds for those categories. To determine thenumber of condition classes, I first categorized thecomposite disturbance gradient into k= 3 to 5groups. I used the 25th and 75th percentiles topartition the disturbance gradient into threedisturbance categories, the 25th, 50th, and 75th
percentiles to partition it into four disturbance
categories, and the 20th, 40th, 60th, and 80thpercentiles to partition it into five disturbancecategories. One-way analysis of variance withmultiple comparisons was used to test whethermean VIBI scores were different amongdisturbance categories and whether means for
-
7/28/2019 A Vegetation Index of Biotic
19/69
VIBI scoring thresholds that best predictedmembership to disturbance classes was identifiedusing classification trees. Given a dataset withpredefined groups, classification trees recursivelypartition that dataset into increasingly homogenoussubsets with regard to the groups (Breiman et al.1984, Urban 2002). At each partition, the treealgorithm identifies the scoring threshold for thepredictor variable that best predicts groupmembership. This process continues until aminimum node size is met. Classification tree
analysis was implemented using Therneau andAtkinsons (2005) rpart package for R software.Minimum node size to be split was set at 15. Treeoverfitting was controlled with an iterative 10-foldcross-validation procedure. Classification accuracywas evaluated by comparing predicted to actualgroup membership.
Indicator species analysis was used to identify
species that were strongly associated with VIBIcondition categories. Indicator species analysisexamines the frequency of occurrence andabundance of species within groups and assigns agroup indicator value based on the specificity andfidelity of a species to that group (Dufrne andLegendre 1997). Group indicator values rangefrom 0 (no indication of group membership) to 100
(perfect indication). The strength of associationwas tested using a Monte Carlo randomizationprocedure with 10,000 iterations. Species withindicator values >25 and P-values
-
7/28/2019 A Vegetation Index of Biotic
20/69
Spatial Autocorrelation Analysis
Spatial autocorrelation can be broadly defined asa significant positive or negative correlation of the
values of a variable as a function of distance (i.e.,
samples of a variable that are closer together in
space having more similar values than those further
away would be an example of positive spatial
autocorrelation). Spatial autocorrelation is a very
general phenomenon that operates at multiple
scales for most ecological and environmentalvariables, and it is an important functional property
of ecosystems (Legendre 1993). Autocorrelated
data are problematic, however, because they
violate an important assumption of many statistical
tests, that observations of variables are independent
from one another. The presence of positive
autocorrelation between closely spaced
observations distorts many tests and increases the
likelihood of erroneous findings of statistical
significance (Legendre and Legendre 1998). This
has been observed for tests of normality (Dutilleul
and Legendre 1992), analysis of variance
(Legendre et al. 1990), and linear regression (Cliff
and Ord 1981). However, Legendre et al. (2002)
have shown that tests of significance for
correlation and regression coefficients were valid
unless both the response and predictor variables
were spatially autocorrelated.
I used two approaches to test for the presence
of spatial autocorrelation. For environmental
variables and derived vegetation variables
(metrics), spatial autocorrelation was evaluated for
each factor independently. Two statistics, Morans
Iand Gearys c, were calculated using Rookcase
software (Sawada 1999). These statistics are
sensitive to departures from normality, and data
were transformed as needed as previously
described. Distances between sites were
calculated from site coordinates projected in
Euclidean space (Montana State Plane, 1983 North
American Datum). Inter-site distances were
divided into 10 classes and values forIand c were
calculated for each class. The number of distance
classes was chosen using Sturges rule based on 30
samples and 435 pairwise comparisons (number of
classes = 1 + 3.3log10(435) = 9.7) (Legendre andLegendre 1998). The significance of correlation
coefficients was tested using a Monte Carlo
randomization procedure with 10,000 iterations.
Because the significance of coefficients was tested
multiple times (once for each distance class),
significance levels were adjusted with a Bonferroni
correction. As the study area was relatively
environmentally homogenous, second-orderstationarity was assumed.
Spatial structure of the entire vegetation
community was examined with a multivariate
Mantel correlogram. Using the method described
by Legendre and Legendre (1998), based on Oden
and Sokal (1986), standardized Mantel statistics
were calculated for a multivariate species distance
matrix (calculated with the Kulczynski distancemeasure) and model matrix based on inter-site
distances. Mantel statistics were calculated for
each distance class and significance values were
calculated by Monte Carlo permutations with 9,999
iterations using PC-ORD. Because of multiple
testing, significance values were corrected with a
Bonferroni procedure.
-
7/28/2019 A Vegetation Index of Biotic
21/69
Human Disturbance Gradient
The composite disturbance gradient was
calculated from a PCA of four variables: AUM,
amount bare ground, bank stability, and browse
intensity. The first principal component explained
58.8% of the variation in the data. It was
considered interpretable as it explained more
variation in the data than expected by chance.
Subsequent principal components did not meet this
criterion. The component was rescaled so that it
ranged between [0, 1], with the least disturbed site
scoring 0 and the most disturbed site scoring 1, and
was used to represent a composite human
disturbance gradient for metric development.
Table 3 shows the contributions of the original
variables to the composite disturbance index.
A PCA including road density was also run. Itwas rejected in favor of the four variable model
because the addition of road density weakened the
interpretability of the first principal component
(component explained 46.9% of the variation, not
much more than that expected by chance) while
road density explained less than 1% of the variation
of the component.
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Proper Functioning Condition Category
surance
ra
en
PFC FAR NF
Table 3. Contribution of individual disturbance
factors to a composite disturbance measure
extracted by principal components analysis.
Measures of pug and hummock density were not
included in the composite human disturbance
gradient The relationship of these measures to
RESULTS
Factor
Variance Explained
(R2)
AUM 0.223
bare ground 0.346
bank stability 0.252browse intensity 0.179
hummocking development. Although the
relationship between pugging and hummocking andsoil texture is only anecdotal for this dataset, there
was a significant correlation between the elevation
of the greenline relative to bankfull discharge and
hummock density (rs= -0.42, P = 0.02) and mean
hummock depth (rs= -0.42, P = 0.02) and a weak
correlation between greenline elevation and pug
density (rs= -0.34, P = 0.07).
The composite disturbance gradient and PFCcategories were positively associated, both for all
sites (F2, 27
= 9.81, P = 0.0006) and BLM sites (F2,
20= 11.81, P = 0.0004). However, while composite
disturbance scores were significantly different
between functioning and functioning at risk
categories (all sites, P = 0.003; BLM sites, P =
0.001) and between functioning and nonfunctioning
categories (all sites, P = 0.002; BLM sites, P =0.001), composite disturbance scores were not
different between functioning at risk and
nonfunctioning categories (all sites, P = 0.81; BLM
sites, P = 0.93) (Figure 3, results from all sites
analysis shown).
-
7/28/2019 A Vegetation Index of Biotic
22/69
-
7/28/2019 A Vegetation Index of Biotic
23/69
Table 5. Formulas used to score metrics. Maximum and percentile values are rounded to nearest percent (relative
cover metrics), hundredth (seedling density), or tenth (cover-weighted means). q0.95
and q0.05
refer to the 95th and 5th
percentiles, respectively.
ValueMetric Maximum 95
thpercentile 5
thpercentile Formula
relative cover of native graminoids 50 %ngram / q0.95relative cover of exotic species 55 5 (max - %exotic) / (max - q0.05)relative cover of annuals/biennials
a 18 0 (max - %ann) / (max - q0.05)willow seedling density (# / m
2)
b0.58 sden / q0.95
cover seedling+young willowsc
9 %yngSalix / q0.95cover-weighted FQAI 30.5 cFQAI / q0.95relative cover of hydrophytes 80 %hydro / q0.95
cover-weighted mean bank stabilityrating 3.4 bank / q0.95
avalues were transformed by log10(%ann + 1) prior to scoring
bvalues were transformed by log10(sden + 0.01) + 2 prior to scoring
cvalues were transformed by log10(%yngSalix + 0.1) + 1 prior to scoring
40
60
80
100
VIBI
Formulas used to compute selected metrics and
metric values for the 95th or 5th percentiles are
shown in Table 5; Figure 4 (facing pages) displays
the discriminatory power and relationships of
selected metrics to the composite disturbance
gradient.
VIBI
The VIBI showed a highly significant responseto the composite disturbance gradient (VIBI =
85.08 47.14 [disturbance score], F1, 28
= 34.32,
R2 = 0.55, P = 0.000003; Figure 5). However,
-
7/28/2019 A Vegetation Index of Biotic
24/69
Least Most
10
20
30
405
0
60
70
%n
ativegram
inoids
0.0 0.2 0.4 0.6 0.8 1.0
10
20
30
405
0
60
70
Least Most
10
20
30
40
50
%e
xoticspecies
0.0 0.2 0.4 0.6 0.8 1.0
10
20
30
40
50
Least Most
0.0
0.2
0.4
0.6
0.8
1.0
1.2
log(%a
nnual/biennial)
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Least Most
0.0
0.5
1.0
1.5
2.0
log(%willowseedlin
gdensity)
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.5
1.0
1.5
2.0
-
7/28/2019 A Vegetation Index of Biotic
25/69
Least Most
0.0
0.5
1.0
1.5
2.0
log(%y
oung
willows)
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.5
1.0
1.5
2.0
Least Most
15
20
25
30
cover-weightedFQAI
0.0 0.2 0.4 0.6 0.8 1.0
15
20
25
30
Least Most
30
40
50
60
70
80
90
%h
ydrophytes
0.0 0.2 0.4 0.6 0.8 1.0
30
40
50
60
70
80
90
Least Most
2.0
2.5
3.0
3.5
cover-weightedbanks
tabilityrating
0 0 0 2 0 4 0 6 0 8 1 0
2.0
2.5
3.0
3.5
-
7/28/2019 A Vegetation Index of Biotic
26/69
Table 6. Accuracy assessment of VIBI scoring thresholds with regard to disturbance classes.
VIBI scores could reliably only differentiate three
condition classes (F2, 27
= 23.09, P = 0.0000001, all
pairwise comparisons significant at the 0.001 level
after Bonferroni correction). Overall, the VIBI
was relatively robust in its ability to differentiate
between these classes (Table 6). While analyses
of variance of the four- and five-category partitions
of the composite disturbance gradient were
significant, VIBI means were not strongly
differentiated among all disturbance categories.
Sites with VIBI scores above 70 were considered
to be reference condition (Figure 6), sites with
scores from 48 to 70 were considered to be
moderately impaired (Figure 7), and sites with
scores below 48 were considered to be severely
impaired (Figure 8). Figure 9 graphically displays
VIBI scoring thresholds, condition classes, and
misclassified cases. Species indicative of eachcondition class are shown in Table 7.
Figure 7. Moderately impaired site; channel shows
evidence of past incisement but is stable.
Figure 8. Severely impaired site; note incised and
Predicted disturbance class
Actual class Least disturbed Moderately disturbed Most disturbed Actual total
Least disturbed 8 0 0 8Moderately disturbed 1 12 2 15
Most disturbed 0 1 6 7
Predicted total 9 13 8 30
Overall accuracy 87%
-
7/28/2019 A Vegetation Index of Biotic
27/69
Figure 9. (A) Tree diagram showing VIBI scoring thresholds associated with disturbance categories. Least = least
disturbed, Moderate = moderately disturbed, Most = most disturbed. (B) Scatterplot of the composite disturbancegradient and VIBI. Symbols represent disturbance categores: = least disturbed sites, = moderately disturbed
sites, = most disturbed sites. Colors represent VIBI classes: green = reference condition, blue = moderately
impaired, red = severely impaired.
|VIBI>=70.25
VIBI>=48.44Least
Moderate Most
0.0 0.2 0.4 0.6 0.8 1.0
20
4
0
60
80
100
Disturbance Gradient
VIBI
Table 7. Species indicative of reference, moderately disturbed, and severely disturbed sites. Indicator valuerepresents the strength of indication (0 = no indication, 100 = perfect indication). P-values were calculated with a
Monte Carlo permutation test. Species with indicator values >25 and P
-
7/28/2019 A Vegetation Index of Biotic
28/69
Whole Community
Relationships among sample units are graphicallydisplayed in Figure 10, which shows the resultsfrom the NMS ordination (three-dimensionalsolution, stress = 12.56, instability
-
7/28/2019 A Vegetation Index of Biotic
29/69
western Montana wetlands. Although not strictly a
functional classification, the concept of floristicquality, which is based on the fidelity of plantspecies to high-integrity habitats, can be used as abroadly integrative measure of site condition. It isespecially pertinent for measuring human-associated stresses, as the tolerance of plantspecies to anthropogenic disturbance is an implicitcriterion in the assignment ofC-values. The utility
of the floristic quality assessment index as avegetation metric has been demonstrated in diversewetland settings (Lopez and Fennessy 2002,DeKeyser et al. 2003, Cohen et al. 2004).
The FQAI has been criticized for the subjectiveassignment ofC-values. In a study of prairiepotholes using C-values assigned by expert opinionfor the Dakotas (Northern Great Plains Floristic
Quality Assessment Panel 2001), Mushet et al.(2002) found that subjectively assigned C-valueswere good indicators of species response and gavecomparable results to C-values that had beenobjectively derived. Although the C-values used inthis study have not been independently verified,they are likely to be similarly robust.
One surprising finding was the relatively poorperformance of the floristic quality assessmentindex, at least as traditionally calculated. TheFQAI is usually computed based on speciespresence/absence data, and this approach has beenfound to be a good indicator of site condition(Lopez and Fennessy 2002, Cohen et al. 2004).However, in this study, the species richness-basedFQAI exhibited a weak correlation withdisturbance. Including exotic species in the
richness-based FQAI provided a marginalimprovement. In contrast, the FQAI weighted byeach species relative cover was stronglycorrelated with the disturbance gradient. This is incontrast to Cohen et al. (2004) who found noi t i FQAI f h i
DISCUSSIONThe goal of this study was to find attributes of
the riparian vegetation community that respondedpredictably to human disturbance and could be usedto assess site condition. Eight such attributes wereidentified and combined into a vegetation index ofbiotic integrity. Overall, this multimetric indexdemonstrated a robust response to grazing-relatedstressors, and VIBI scores could be used toclassify a site into one of three disturbance
categories with relatively high accuracy. Within thereference domain considered small-ordermontane streams able to support woody vegetation the VIBI appears to be a good indicator of sitecondition. This is consistent with other multimetricvegetation studies that have found plants to be goodindicators of wetland and riparian condition(DeKeyser et al. 2003, Mack 2004, Jones 2004,
Ferreira et al. 2005).Both a strength and complication of usingvegetation as an indicator of site condition is thelarge number of species often involved. Forexample, 178 species of vascular plants weresampled in the course of this study, and the meanrichness was 43 7 species per site. A strength ofthe multimetric approach is that species aregrouped by the expected similarity of their responseto disturbance or stress. This makes use ofredundancies in species responses within groupsand can thereby reduce the noise often generatedwhen the response of all species is consideredsimultaneously. This study made use of speciesgroups based on functionality, taxonomy, andnativity. The utility of vegetation classificationsbased on common attributes, adaptations, or
responses of species to environmental factors, haslong been recognized (Raunkiaer 1934, Grime 1977,Grime 1988, Lavorel and Garnier 2002, Pausas andLavorel 2003). Functional groups in particular havebeen shown to be an effective approach to
l ti t ti t i l t d
-
7/28/2019 A Vegetation Index of Biotic
30/69
were used to establish a broad disturbance gradientfor sampling purposes. However, PFCassessments may not adequately differentiate
between moderately and highly disturbed sites, atleast when grazing is the primary stressor. This isevidenced by the lack of difference in mean PCA-derived disturbance scores between the functioningat risk and nonfunctioning categories. This lack ofassociation may reflect in part the differentpurposes of these measures of disturbance: thecomposite disturbance gradient was constructed by
finding linear combinations of variables that wereexpected to measure different aspects of grazing-associated stresses, while the PFC is a moregeneral method to evaluate site condition.
Another aspect of the site selection processshould be reconsidered: in defining the siteselection criteria for this study, the samplinguniverse was restricted to sites able to support tall
woody vegetation (i.e., willows). Site potential wasverified by either previous BLM surveys, whichcharacterized sites by Hansen et al.s (1995)vegetation community classification or by review ofU.S. Geological Survey digital orthophoto imagery.This was done to focus on the most typical streamreaches (which do support woody vegetation) andto reduce environmental heterogeneity by excluding
forested, sagebrush, or herbaceous-dominatedstream reaches (i.e., sedge meadows). Althoughreducing environmental heterogeneity is animportant design consideration when developingmultimetric indices (Teels and Adamus 2002), anunfortunate result of this stratification was thepotential undersampling of extremely disturbed siteswhere grazing had completely removed woody
cover. All the sites sampled in this study, even themost heavily disturbed, supported willow cover,although at heavily disturbed sites this cover wasusually exclusively provided by mature orsenescent willows.
Another improvement would be to develop a
repens L.), and common dandelion (Taraxacumofficinale G.H. Weber ex Wiggers).
Although in this study the FQAI was used as a
component in a multimetric index, floristic qualityassessments should have broader applicability. Inassigning C-values, the expert panel was not limitedto the species sampled in this study but consideredall species likely to occur in western Montanawetlands (the species list was taken from Lesicaand Husby (2001, Appendix A)). Thus, the FQAIand related measures of floristic quality can be
tested and applied as a stand-alone indicator of sitecondition to all wetland types in western Montana,not just the limited subset considered here. Furthertesting should be done to compare the relativeutility of the presence-absence and cover-weightedformulations of the FQAI.
Recommendations for Future
Improvements
An important next step is to validate the VIBIand to expand its applicability. This study examinedvegetation response to a single, albeit complex,stressor. The VIBI should be validated atadditional environmentally similar sites where
grazing is the primary human stressor. However, tobe broadly applicable, the VIBI will need to begeneralized so that it is responsive to otheranthropogenic stressors, especially those thatmodify hydrology. Some applicability of the VIBIas formulated here should be expected, as one ofthe effects of overgrazing can be bank erosion andstream channel downcutting, which can affecthydrology and make a site drier. Functionally,there may be some overlap between grazing-induced stresses on the vegetation community andother stressors that cause hydrologic alterations.Several of the metrics developed here, includingwillow seedling density, absolute cover of willow
-
7/28/2019 A Vegetation Index of Biotic
31/69
that the specific results are idiosyncratic to thecollected dataset. A next step would either be tomodel the composite disturbance gradient (e.g.,
generalize the results of the PCA by findingexplanatory equations) or to develop a rule-basedprocedure. Lopez and Fennessy (2002) used arule-based approach to describe wetlanddisturbance: wetlands were ordered into one of 24categories based on buffer conditions and presenceof hydrologic modifications. Ohio EPA used theirrapid assessment method as a measure of site
disturbance (Mack et al. 2000). (This lastapproach would be somewhat circular, as the VIBIis meant to be used to validate DEQs rapidassessment.) Developing a more generalizeddisturbance measure will become more of an issueas the VIBIs reference domain broadens toinclude greater environmental and anthropogenicheterogeneity.
A parallel issue is to limit disturbance factors tovariables that are measurable at all sites. Forexample, three of the disturbance factors usedhere, amount bare ground, bank stability, andbrowse intensity, were measured on-site. Thefourth, livestock use (AUM), was readily availableonly because sites were sampled on public land.
Therefore, AUM is not likely to be easilygeneralized and should probably be removed fromfuture studies.
Finally, there is the question of improving thebroader utility of the VIBI. As the VIBI issufficiently validated (and possibly modified), it willbecome a useful tool to assess riparian areacondition and will provide validation for rapidassessments. However, because many of themetrics require the entire vegetation community tobe enumerated, the VIBI requires extensive
botanical expertise and time. An ongoing goal inrefining the VIBI should be to use more easilymeasured metrics that can perhaps be ultimatelyincorporated into the rapid assessment method.Current examples are willow seedling density andcover of young and seedling willows. Cohen et al.(2005) used classification and regression trees toidentify indicator species for different wetland
condition categories, thereby lessening the botanicalexpertise needed to assess wetland condition.Likewise, the indicator species recognized heremay form the basis of identifying key plant speciesthat are consistently associated with certain levelsof disturbance.
-
7/28/2019 A Vegetation Index of Biotic
32/69
LITERATURE CITEDAbernathy, B. and I. D. Rutherfurd. 2001. The
distribution and strength of riparian tree roots inrelation to riverbank reinforcement.Hydrological Processes 15:63-79.
Andreas, B. K., J. J. Mack, and J. S. McCormac.2004. Floristic quality assessment index(FQAI) for vascular plants and mosses for theState of Ohio. Ohio Environmental ProtectionAgency, Division of Surface Water, Wetland
Ecology Group, Columbus, Ohio.Apfelbeck, R. S. 2001. Development of biocriteriafor wetlands in Montana. Pages 141-166 in R.B. Rader, D. P. Batzer, and S. A. Wissinger,editors. Bioassessment and management ofNorth American freshwater wetlands. JohnWiley & Sons, Inc., New York, New York.
Barbour, M. T., J. Gerritsen, G. E. Griffith, R.Frydenborg, E. McCarron, J. S. White, and M.L. Bastian. 1996. A framework for biologicalcriteria for Florida streams using benthicmacroinvertebrates. Journal of the NorthAmerican Benthological Society 15:185-211.
Barbour, M. T., J. Gerritsen, B. D. Snyder, and J.B. Stribling. 1999. Rapid bioassessmentprotocols for use in streams and wadeablerivers: periphyton, benthic macroinvertebratesand fish, Second Edition. EPA 841-B-99-002,U.S. Environmental Protection Agency, Officeof Water, Washington, D.C.
Bengeyfield, P. 1999. Analysis of reference reachdata and the prediction of stream types. Pages245-254 in D. S. Olsen and J. P. Potyondy,editors. Proceedings of the SpecialtyConference on Wildland Hydrology (Bozeman,
Montana). American Water ResourcesAssociation, Hendon, Virginia.
Blocksom, K. A. 2003. A performance comparisonof metric scoring methods for a multimetricindex for mid-Atlantic highland streams.E i t lM t 31 670 682
Borth, C. S. 1998. Effects of land use on
vegetation in glaciated depressional wetlands inwestern Montana. Masters Thesis. MontanaState University, Bozeman, Montana.
Box, G. E. P. and D. R. Cox. 1964. An analysis oftransformations. Journal of the Royal StatisticalSociety B 26:211-243.
Breiman, L., J. H. Friedman, R. A. Olshen, and C.J. Stone. 1984. Classification and regression
trees. Chapman & Hall, New York, NewYork.Brinson, M. M., B. L. Swift, R. C. Plantico, and J.S. Barclay. 1981. Riparian ecosystems: theirecology and status. U.S. Fish and WildlifeService Biological Report 81, U.S. GovernmentPrinting Office, Washington, D.C.
Brookshire, E. N. J., J. B. Kauffman, D. Lytjen,and N. Otting. 2002. Cumulative effects ofwild ungulate and livestock herbivory onriparian willows. Oecologia 132:559-566.
Bryce, S. A., R. M. Hughes, and P. R. Kaufmann.2002. Development of a bird integrity index:using bird assemblages as indicators of ripariancondition. Environmental Management 30:294-310.
Clary, W. P. 1999. Stream channel and vegetationresponse to late spring cattle grazing. Journalof Range Management 52:218-227.
Clary, W. P. and J. W. Kinney. 2002. Streambankand vegetation response to simulated cattlegrazing. Wetlands 22:139-148.
Cliff, A. D. and J. K. Ord. 1981. Spatial processes- models and applications. Pion, London.
Cohen, M. J., S. Carstenn, and C. R. Lane. 2004.Floristic quality indices for biotic assessment of
depressional marsh condition in Florida.Ecology 14:784-794.
Cohen, M. J., C. R. Lane, K. C. Reiss, J. A.Surdick, E. Bardi, and M. T. Brown. 2005.Vegetation based classification trees for rapid
t f i l t d tl d diti
-
7/28/2019 A Vegetation Index of Biotic
33/69
Cooper, S. V. 2004. Annotated bibliography ofquantitative sampling methods for watershedassessment. Montana Natural Heritage
Program, Helena, Montana.Crowe, E. and R. Clausnitzer. 1997. Mid-montane
wetland plant associations of the Malheur,Umatilla, and Wallowa-Whitman NationalForests. Technical Paper R6-NR-ECOL-TP-22-97, USDA Forest Service, PacificNorthwest Region.
Dahl, T. E. 1990. Wetland losses in the United
States 1780s to 1980s. U.S. Department ofInterior, Fish and Wildlife Service, Washington,D.C.
Danielson, T. J. 2002. Methods for evaluatingwetland condition: introduction to wetlandbiological assessment. EPA-822-R-02-014,Office of Water, U.S. EnvironmentalProtection Agency, Washington, D.C.
Daubenmire, R. F. 1959. A canopy-coveragemethod of vegetation analysis. NorthwestScience 33:43-64.
DeKeyser, E. S., D. R. Kirby, and M. J. Ell. 2003.An index of plant community integrity:development of the methodology for assessingprairie wetland plant communities. EcologicalIndicators 3:119-133.
Dufrne, M. and P. Legendre. 1997. Speciesassemblages and indicator species: the need fora flexible asymmetrical approach. EcologicalMonographs 67:345-366.
Dutilleul, P. and P. Legendre. 1992. Lack ofrobustness in two tests of normality againstautocorrelation in sample data. Journal ofStatistical Computation and Simulation 42:79-91.
Fennessy, S., M. Gernes, J. Mack, and D. H.Wardrop. 2002. Methods for evaluatingwetland condition: using vegetation to assessenvironmental conditions in wetlands. EPA-822-R-02-020, Office of Water, U.S.
Finch, D. M. and L. F. Ruggiero. 1993. Wildlife andbiological diversity in the Rocky Mountains andnorthern Great Plains. Natural Areas Journal
13:191-203.Flenniken, M., R. R. McEldowney, W. C.
Leininger, G. W. Frasier, and M. J. Trlica. 2001.Hydrologic responses of a montane riparianecosystem following cattle use. Journal ofRange Management 54:567-574.
Fore, L. S. and C. Grafe. 2002. Using diatoms toassess the biological condition of large rivers in
Idaho (USA). Freshwater Biology 47:2015-2037.Fox, J. 2005. car: Companion to applied regression
package for R, version 1.0-16. Available athttp://cran.r-project.org/.
Francis, C. M., M. J. W. Austen, J. M. Bowles,and W. B. Draper. 2000. Assessing floristicquality in southern Ontario woodlands. Natural
Areas Journal 20:66-77.Friedel, M. H., G. N. Bastin, and G. F. Griffin.1988. Range assessment and monitoring in aridlands: the derivation of functional groups tosimplify vegetation data. Journal ofEnvironmental Management 27:85-97.
Frontier, S. 1976. tude de la dcroissance desvaleurs propres dans une analyse en
composantes principales: comparaison avec lemodle du bton bris. Journal of ExperimentalMarine Biology and Ecology 25:67-75.
Green, D. M. and J. B. Kauffman. 1995.Succession and livestock grazing in anortheastern Oregon riparian ecosystem.Journal of Range Management 48:307-313.
Grime, J. P. 1977. Evidence for the existence ofthree primary strategies in plants and itsrelevance to ecological and evolutionary theory.American Naturalist 111:1169-1194.
Grime, J. P. 1988. The C-S-R model of primaryplant strategies - origins, implications and tests.Pages 371-393 in L. D. Gottlieb and S. K. Jain,
-
7/28/2019 A Vegetation Index of Biotic
34/69
Helgen, J. C. and M. C. Gernes. 2001. Monitoringthe condition of wetlands: indexes of biologicalintegrity using invertebrates and vegetation.
Pages 167-185 in R. B. Rader, D. P. Batzer,and S. A. Wissinger, editors. Bioassessmentand management of North Americanfreshwater wetlands. John Wiley & Sons, Inc.,New York, New York.
Hughes, R. M., P. R. Kaufmann, A. T. Herlihy, T.M. Kincaid, L. Reynolds, and D. P. Larsen.1998. A process for developing and evaluating
indices of fish assemblage integrity. CanadianJournal Fisheries and Aquatic Sciences55:1618-1631.
Hyndman, R. J. and Y. Fan. 1996. Sample quantilesin statistical packages. American Statistician50:361-365.
Ihaka, R. and R. Gentleman. 1996. R: a languagefor data analysis and graphics. Journal of
Computational and Graphical Statistics 5:299-314.Jackson, D. A. 1993. Stopping rules in principal
components analysis: a comparison ofheuristical and statistical approaches. Ecology74:2204-2214.
Jones, W. M. 2004. Using vegetation to assesswetland condition: a multimetric approach for
temporarily and seasonally flooded depressionalwetlands and herbaceous-dominatedintermittent and ephemeral riverine wetlands inthe northwestern glaciated plains ecoregion,Montana. Report to the Montana Departmentof Environmental Quality and U.S.Environmental Protection Agency, MontanaNatural Heritage Program, Helena, Montana.
Karr, J. R. 1981. Assessment of biotic integrityusing fish communities. Fisheries 6:21-27.
Karr, J. R. and E. W. Chu. 1999. Restoring life inrunning waters: better biological monitoring.Island Press, Washington, D.C.
Kauffman, J. B. and W. C. Krueger. 1984.
Kauffman, J. B., W. C. Krueger, and M. Vavra.1983b. Impact of cattle on streambanks innortheastern Oregon. Journal of Range
Management 36:683-685.Kauffman, J. B., A. S. Thorpe, and E. N. J.
Brookshire. 2004. Livestock exclusion andbelowground ecosystem responses in riparianmeadows of eastern Oregon. EcologicalApplications 14:1671-1679.
Keddy, P. A. 2000. Wetland ecology: principles andconservation. Cambridge University Press,
Cambridge.Keigley, R. B. and M. R. Frisina. 1998. Browseevaluation by analysis of growth form. Volume1, methods for evaluating condition and trend.Montana Department of Fish, Wildlife & Parks,Helena, Montana.
Kimberling, D. N., J. R. Karr, and L. S. Fore.2001. Measuring human disturbance using
terrestrial invertebrates in the shrub-steppe ofeastern Washington (USA). EcologicalIndicators 1:63-81.
King, R. S., M. E. Baker, D. F. Whigham, D. E.Weller, T. E. Jordan, P. F. Kazyak, and M. K.Hurd. 2005. Spatial considerations for linkingwatershed land cover to ecological indicators instreams. Ecological Applications15:137-153.
Klemm, D. J., K. A. Blocksom, F. A. Fulk, A. T.Herlihy, R. M. Hughes, P. R. Kaufmann, D. V.Peck, J. L. Stoddard, W. T. Theony, and M. B.Griffith. 2003. Development and evaluation ofa macroinvertebrate biotic integrity index(MBII) for regionally assessing mid-Atlantichighland streams. Environmental Management31:656-669.
Kruskal, J. B. 1964. Nonmetric multidimensionalscaling: a numerical method. Psychometrika29:115-129.
Landsberg, J., S. Lavorel, and J. Stol. 1999.Grazing response groups among understoreyplants in arid rangelands. Journal of Vegetation
-
7/28/2019 A Vegetation Index of Biotic
35/69
Lavorel, S., S. McIntyre, J. Landsberg, and T. D.A. Forbes. 1997. Plant functionalclassifications: from general groups to specific
groups based on response to disturbance.Trends in Ecology and Evolution 12:474-478.
Legendre, P. 1993. Spatial autocorrelation: troubleor new paradigm? Ecology 74:1659-1673.
Legendre, P., M. R. T. Dale, M.-J. Fortin, J.Gurevitch, M. Hohn, and D. Myers. 2002. Theconsequences of spatial structure for thedesign and analysis of ecological field surveys.
Ecography 25:601-615.Legendre, P. and L. Legendre. 1998. Numericalecology, 2nd English edition. Elsevier ScientificPublishing Company, Amsterdam, TheNetherlands.
Legendre, P., N. L. Oden, R. R. Sokal, A. Vaudor,and J. Kim. 1990. Approximate analysis ofvariance of spatially autocorrelated regional
data. Journal of Classification 7:53-75.Lesica, P. and P. Husby. 2001. Field guide toMontanas wetland vascular plants. MontanaWetlands Trust, Helena, Montana.
Levene, H. 1960. Robust tests for equality ofvariance. Pages 278-292 in I. Olkin, editor.Contributions to probability and statistics.Stanford University Press, Palo Alto,
California.Lonn, J. D., B. Skipp, E. T. Ruppel, S. U. Janecke,W. J. Perry, J. W. Sears, M. J. Bartholomew,M. C. Stickney, W. J. Fritz, H. A. Hurlow, andR. C. Thomas. 2000. Preliminary geologic mapof the Lima 30' x 60' quadrangle, southwestMontana. Open File Report MBMG 408,Montana Bureau of Mines and Geology, Butte,Montana.
Lopez, R. D. and M. S. Fennessy. 2002. Testingthe floristic quality assessment index as anindicator of wetland condition. EcologicalApplications 12:487-97.
Mack, J. J., M. Micacchion, L. D. Augusta, and G.
Mack, J. J. 2004. Integrated wetland assessmentprogram, part 4: vegetation index of bioticintegrity (VIBI) and tiered aquatic life uses
(TALUs) for Ohio wetlands. Ohio EPATechnical Report WET/2004-4, OhioEnvironmental Protection Agency, WetlandEcology Group, Division of Surface Water,Columbus, Ohio.
Mantel, N. 1967. The detection of diseaseclustering and a generalized regressionapproach. Cancer Research 27:209-220.
Martin, D. W. and J. C. Chambers. 2001. Effectsof water table, clipping, and speciesinteractions on Carex nebrascensis and Poapratensis in riparian meadows. Wetlands21:422-430.
Mather, P. M. 1976. Computational methods ofmultivariate analysis in physical geography. J.Wiley & Sons, London.
McCormick, F. H., R. M. Hughes, P. R.Kaufmann, D. V. Peck, J. L. Stoddard, and A.T. Herlihy. 2001. Development of an index ofbiotic integrity for the mid-Atlantic highlandsregion. Transactions of the American FisheriesSociety 130:857-877.
McCune, B. and M. J. Mefford. 1999. PC-ORD.Multivariate analysis of ecological data, version
4.36. MjM Software Design, Gleneden Beach,Oregon.McCune, B. and J. B. Grace. 2002. Analysis of
ecological communities. MjM Software Design,Gleneden Beach, Oregon.
McIntyre, S., S. Lavorel, and R. M. Tremont. 1995.Plant life-history attributes: their relationship todisturbance response in herbaceous vegetation.Journal of Ecology 83:31-44.
Mebane, C. A., T. R. Maret, and R. M. Hughes.2003. A performance comparison of metricscoring methods for a multimetric index formid-Atlantic highlands streams. Transactions ofthe American Fisheries Society 132:239-261.
-
7/28/2019 A Vegetation Index of Biotic
36/69
-
7/28/2019 A Vegetation Index of Biotic
37/69
Swink, F. A. and G. S. Wilhelm. 1979. Plants of theChicago region: a checklist of the vascularflora of the Chicago region, with keys, notes on
local distribution, ecology, and taxonomy, and asystem for evaluation of plant communities.Morton Arboretum, Lisle, Illinois.
Teels, B. M. and P. Adamus. 2002. Methods forevaluating wetland condition: developingmetrics and indices of biological integrity. EPA-822-R-02-016, Office of Water, U.S.Environmental Protection Agency, Washington,
D.C.Therneau, T. M. and B. Atkinson. 2005. rpart:Recursive partitioning package for R, version3.1-23. Available at http://cran.r-project.org/.
Thorne, M. S., P. J. Meiman, Q. D. Skinner, M. A.Smith, and J. L. Dodd. 2005. Clippingfrequency affects canopy volume and biomassproduction in planeleaf willow (Salix
planifolia var. planifolia Prush). RangelandEcology and Management 58:41-50.Urban, D. L. 2002. Classification and regression
trees. Pages 222-232 in B. McCune and J. B.Grace, editors. Analysis of ecologicalcommunities. MjM Software Design, GlenedenBeach, Oregon.
US Census Bureau. 2003. Montana roads from
TIGER/line files. U.S. Census BureauGeography Division, Helena, Montana.Available at http://nris.state.mt.us/nsdi/nris/road2000/road2000.html.
US Geological Survey. 2002. National elevationdataset for Montana. U.S. Geological Survey,Helena, Montana. Available at http://nris.state.mt.us/nsdi/nris/el10/dems.html.
USEPA. 1998. Lake and reservoir bioassessmentand biocriteria. Technical Guidance DocumentEPA 841-B-98-007, U.S. EnvironmentalProtection Agency, Office of Wetlands,Oceans, and Watersheds, Washington, D.C.
Venables, W. N. and B. D. Ripley. 2002. Modernapplied statistics with S, 4th edition. Springer-Verlag, New York, New York.
Western Regional Climate Center. 2005. Montanaclimate summaries, National Climatic DataCenter 1971-2000 climate normals. NationalOceanic and Atmospheric Administration.Available at http://www.wrcc.dri.edu/summary/climsmmt.html. Accessed on 2/21/2005.
Wheeler, M. A., M. J. Trlica, G. W. Frasier, and J.D. Reeder. 2002. Seasonal grazing affects soil
physical properties of a montane ripariancommunity. Journal of Range Management55:49-56.
Whittaker, R. J. 1972. Evolution and measurementof species diversity. Taxon 21:213-51.
Wilcox, D. A., J. E. Meeker, P. L. Hudson, B. J.Armitage, M. G. Black, and D. G. Uzarski.2002. Hydrologic variability and the application
of index of biotic integrity metrics to wetlands:a Great Lakes evaluation. Wetlands 22:588-615.
Wilhelm, G. and D. Ladd. 1988. Natural areaassessment in the Chicago region. Pages 361-375 in Transactions of the 53rd NorthAmerican Wildlife and Natural ResourcesConference (Louisville, Kentucky). Wildlife
Management Institute, Washington, D.C.Winward, A. H. 2000. Monitoring the vegetationresources in riparian areas. General TechnicalReport RMRS-GTR-47, U.S. Department ofAgriculture, Forest Service, Rocky MountainResearch Station, Ogden, Utah.
Woods, A. J., J. M. Omernik, J. A. Nesser, J.Shelden, and S. H. Azevedo. 1999. Ecoregionsof Montana. (1:1,500,000 map), U.S.Geological Survey, Reston, Virginia.
-
7/28/2019 A Vegetation Index of Biotic
38/69
-
7/28/2019 A Vegetation Index of Biotic
39/69
APPENDIXA. COEFFICIENTSOFCONSERVATIONFORSELECTEDWET-
LANDPLANTSTHATOCCURINWESTERNMONTANA.
-
7/28/2019 A Vegetation Index of Biotic
40/69
-
7/28/2019 A Vegetation Index of Biotic
41/69
Appendix A. Coefficients of conservation for selected wetland plants that occur in western Montana.
Coefficients of conservatism were assigned to 747 plant species known to occur in wetlands in western
Montana. Species were selected based on the species list in Lesica and Husby (2001, Appendix A).Coefficients for a few additional non-wetland species were assigned because they were sampled in the
course of the study. Coefficients were determined by a panel of expert botanists. Panel members were
Stephen Cooper (Vegetation Ecologist, Montana Natural Heritage Program), Marc Jones (Ecologist,
Montana Natural Heritage Program), Peter Lesica (Botanical Consultant), Mary Manning (Vegetation
Ecologist, U.S. Forest Service), Scott Mincemoyer (Botanist, Montana Natural Heritage Program), JohnPierce (Botanical Consultant), and Steve Shelly (Regional Botanist, U.S. Forest Service). Coefficients for
345 species were assigned by the entire committee; coefficients for the remaining 402 species were
assigned by Marc Jones, Peter Lesica, and John Pierce. Nomenclature follows the federal naming
standard (Natural Resources Conservation Service 2004). For unfamiliar names, a partial synonymy can
be found by consulting the PLANTS database at http://plants.usda.gov.
Literature Cited
Lesica, P. and P. Husby. 2001. Field guide to Montana's wetland vascular plants. Montana WetlandsTrust, Helena, Montana.
Natural Resources Conservation Service. 2004. The PLANTS database, version 3.5. U.S. Department ofAgriculture, Natural Resources Conservation Service, National Plant Data Center. Available at
http://plants.usda.gov.
http://plants.usda.gov/http://plants.usda.gov/ -
7/28/2019 A Vegetation Index of Biotic
42/69
Coefficients of conservatism (C) for 747 wetland plants species known to occur in western Montana.
C Scientific Name Common Name
4 Acer glabrum Torr. Rocky Mountain maple
2 Acer negundo L. boxelder6 Achnatherum nelsonii (Scribn.) Barkworth Columbia needlegrass
5 Aconitum columbianum Nutt. Columbian monkshood
7 Actaea rubra (Ait.) Willd. red baneberry
7 Adiantum aleuticum (Rupr.) Paris Aleutian maidenhair
5 Agoseris aurantiaca (Hook.) Greene orange agoseris
4 Agoseris glauca (Pursh) Raf. pale agoseris
3 Agrostis exarata Trin. spike bentgrass
1 Agrostis gigantea Roth redtop
8 Agrostis humilis Vasey alpine bentgrass
2 Agrostis scabra Willd. rough bentgrass
6 Allium brevistylum S. Wats. shortstyle onion
6 Allium schoenoprasum L. wild chives
6 Alnus incana (L.) Moench gray alder
6 Alnus viridis (Vill.) Lam. & DC. green alder
4 Alopecurus aequalis Sobol. shortawn foxtail
6 Alopecurus alpinus Sm. boreal alopecurus
4 Alopecurus carolinianus Walt. Carolina foxtail2 Alopecurus geniculatus L. water foxtail
0 Alopecurus pratensis L. meadow foxtail
0 Amaranthus blitoides S. Wats. mat amaranth
7 Amaranthus californicus (Moq.) S. Wats. California amaranth
3 Ambrosia psilostachya DC. Cuman ragweed
1 Ambrosia trifida L. great ragweed
10 Amerorchis rotundifolia (Banks ex Pursh) Hultn roundleaf orchid
2 Androsace filiformis Retz. filiform rockjasmine6 Anemone parviflora Michx. smallflowered anemone
5 Angelica arguta Nutt. Lyall's angelica
7 Angelica dawsonii S. Wats. Dawson's angelica
4 Angelica pinnata S. Wats. small-leaf angelica
3 Antennaria corymbosa E. Nels. flat-top pussytoes
3 Antennaria microphylla Rydb. littleleaf pussytoes
4 Apocynum cannabinum L. Indianhemp
6 Aquilegia caerulea James Colorado blue columbine
6 Aquilegia formosa Fisch. ex DC. western columbine0 Arenaria serpyllifolia L. thymeleaf sandwort
3 Argentina anserina (L.) Rydb. silverweed cinquefoil
5 Arnica amplexicaulis Nutt. clasping arnica
5 Arnica chamissonis Less. Chamisso arnica
7 A i l if li D C E l f i
-
7/28/2019 A Vegetation Index of Biotic
43/69
C Scientific Name Common Name
4 Artemisia tridentata Nutt. ssp. wyomingensis Beetle &
Young
Wyoming big sagebrush
0 Asclepias speciosa Torr. showy milkweed3 Astragalus agrestis Dougl. ex G. Don purple milkvetch
6 Astragalus americanus (Hook.) M.E. Jones American milkvetch
3 Astragalus canadensis L. Canadian milkvetch
5 Athyrium filix-femina (L.) Roth common ladyfern
0 Atriplex patula L. spear saltbush
5 Atriplex truncata (Torr. ex S. Wats.) Gray wedgescale saltbush
7 Bacopa rotundifolia (Michx.) Wettst. disk waterhyssop
4 Barbarea orthoceras Ledeb. American yellowrocket
0 Barbarea vulgaris Ait. f. garden yellowrocket
4 Beckmannia syzigachne (Steud.) Fern. American sloughgrass
7 Berula erecta (Huds.) Coville cutleaf waterparsnip
8 Betula nana L. dwarf birch
5 Betula occidentalis Hook. water birch
4 Bidens cernua L. nodding beggartick
6 Bidens tripartita L. threelobe beggarticks
6 Bidens vulgata Greene big devils beggartick
6 Botrychium lanceolatum (Gmel.) Angstr. lanceleaf grapefern4 Botrychium lunaria (L.) Sw. common moonwort
8 Botrychium multifidum (Gmel.) Trev. leathery grapefern
7 Botrychium pinnatum St. John northern moonwort
6 Botrychium simplex E. Hitchc. little grapefern
7 Boykinia majorGray large boykinia
6 Bromus ciliatus L. fringed brome
0 Bromus inermis Leyss. smooth brome
5 Bromus marginatus Nees ex Steud. mountain brome5 Calamagrostis canadensis (Michx.) Beauv. bluejoint
6 Calamagrostis stricta (Timm) Koel. slimstem reedgrass
6 Callitriche hermaphroditica L. northern water-starwort
3 Callitriche heterophylla Pursh twoheaded water-starwort
7 Caltha leptosepala DC. white marsh marigold
6 Calypso bulbosa (L.) Oakes fairy slipper
6 Camassia quamash (Pursh) Greene small camas
5 Camissonia subacaulis (Pursh) Raven diffuseflower evening-primrose
7 Campanula parryi Gray Parry's bellflower3 Campanula rotundifolia L. bluebell bellflower
9 Campanula uniflora L. arctic bellflower
6 Canadanthus modestus (Lindl.) Nesom giant mountain aster
7 Cardamine breweri S. Wats. Brewer's bittercress
3 C d i li N tt littl t bitt
-
7/28/2019 A Vegetation Index of Biotic
44/69
-
7/28/2019 A Vegetation Index of Biotic
45/69
C Scientific Name Common Name
4 Carex praticola Rydb. meadow sedge
7 Carex pyrenaica Wahlenb. Pyrenean sedge
7 Carex sartwellii Dewey Sartwell's sedge8 Carex saxatilis L. rock sedge
7 Carex scoparia Schkuhr ex Willd. broom sedge
8 Carex scopulorum Holm mountain sedge
8 Carex simulata Mackenzie analogue sedge
7 Carex spectabilis Dewey showy sedge
8 Carex sprengelii Dewey ex Spreng. Sprengel's sedge
4 Carex stipata Muhl. ex Willd. owlfruit sedge
8 Carex sychnocephala Carey manyhead sedge
10 Carex tenuiflora Wahlenb. sparseflower sedge
9 Carex torreyi Tuckerman Torrey's sedge
3 Carex utriculata Boott Northwest Territory sedge
5 Carex vesicaria L. blister sedge
8 Carex viridula Michx. little green sedge
6 Carex vulpinoidea Michx. fox sedge
4 Castilleja miniata Dougl. ex Hook. giant red Indian paintbrush
3 Castilleja minor(Gray) Gray lesser Indian paintbrush
7 Castilleja occidentalis Torr. western Indian paintbrush7 Castilleja rhexiifolia Rydb. splitleaf Indian paintbrush
7 Castilleja sulphurea Rydb. sulphur Indian paintbrush
3 Catabrosa aquatica (L.) Beauv. water whorlgrass
3 Ceratophyllum demersum L. coon's tail
1 Chamerion angustifolium (L.) Holub fireweed
0 Chenopodium album L. lambsquarters
3 Chenopodium rubrum L. red goosefoot
4 Chrysosplenium tetrandrum (Lund ex Malmgr.) Th. Fries northern golden saxifrage7 Cicuta bulbifera L. bulblet-bearing water hemlock
4 Cicuta douglasii (DC.) Coult. & Rose western water hemlock
3 Cicuta maculata L. spotted water hemlock
5 Circaea alpina L. small enchanter's nightshade
0 Cirsium arvense (L.) Scop. Canada thistle
5 Cirsium scariosum Nutt. meadow thistle
4 Cirsium undulatum (Nutt.) Spreng. wavyleaf thistle
0 Cirsium vulgare (Savi) Ten. bull thistle
2 Claytonia perfoliata Donn ex Willd. miner's lettuce5 Claytonia sibirica L. Siberian springbeauty
6 Coeloglossum viride (L.) Hartman longbract frog orchid
3 Collomia linearis Nutt. tiny trumpet
7 Comarum palustre L. purple marshlocks
0 C i l L i h l k
-
7/28/2019 A Vegetation Index of Biotic
46/69
C Scientific Name Common Name
0 Cynoglossum officinale L. gypsyflower
8 Cyperus schweinitzii Torr. Schweinitz's flatsedge
7 Cypripedium fasciculatum Kellogg ex S. Wats. clustered lady's slipper9 Cypripedium parviflorum Salisb. lesser yellow lady's slipper
10 Cypripedium passerinum Richards. sparrowegg lady's slipper
10 Cystopteris montana (Lam.) Bernh. ex Desv. mountain bladderfern
5 Danthonia intermedia Vasey timber oatgrass
3 Dasiphora floribunda (Pursh) Kartesz, comb. nov. ined. shrubby cinquefoil
7 Delphinium depauperatum Nutt. slim larkspur
6 Delphinium glaucum S. Wats. Sierra larkspur
7 Deschampsia caespitosa (L.) Beauv. tufted hairgrass
5 Deschampsia danthonioides (Trin.) Munro annual hairgrass
4 Deschampsia elongata (Hook.) Munro slender hairgrass
9 Dichanthelium acuminatum (Sw.) Gould & C.A. Clark var.
fasciculatum (Torr.) Freckmann
western panicgrass
5 Distichlis spicata (L.) Greene inland saltgrass
7 Dodecatheon jeffreyi Van Houtte Sierrra shootingstar
5 Dodecatheon pulchellum (Raf.) Merr. darkthroat shootingstar
7 Draba aurea Vahl ex Hornem. golden draba
9 Dryopteris cristata (L.) Gray crested woodfern0 Echinochloa muricata (Beauv.) Fern. rough barnyardgrass
5 Echin